Cameras are commonly used to capture an image of a scene that includes one or more objects. Unfortunately, some of the images are blurred. For example, when the camera is not properly focused, one or more of the objects captured in the image will be blurred. Depending upon the degree of the improper focus, the resulting image may not be satisfactory.
There exists a number of deconvolution methods for reducing defocus blur in a blurry image. Typically, these methods require that a blur kernel size be estimated. Unfortunately, it can be very difficult to accurately automatically estimate the size of the blur kernel.